Desarrollo de un modelo predictivo utilizando técnicas de aprendizaje supervisado para detectar la moniliasis en plantas de cacao de la Provincia de Orellana

Abstract

The answer to the moniliasis enigma lies in science and technology with the project developed in Orellana Province, where moniliasis is a fungal disease that causes significant losses to farmers. Moniliasis severely affects cocoa crops, and it is difficult to detect its presence early.  Data from sensors and manual records were collected to train and validate a predictive model using supervised learning, where environmental conditions and disease symptoms were analysed. Design science methodology was applied based on three cycles: the relevance, rigour and design cycle. In the relevance cycle the problem and the need for the model were defined, in the rigour cycle a preliminary investigation was carried out to determine the feasibility of the objective and finally in the design cycle the data was modelled with machine learning algorithms and the prediction model was implemented and tested to verify its correct functioning.

The model was shared with cocoa farming families in Orellana, demonstrating its effectiveness. This will allow farmers to take appropriate and timely control measures to prevent the spread of the disease and thus increase cocoa production and quality.